Subject: Unable to Mount Root Partition (NFS)
To: None <netbsd-help@netbsd.org>
From: Jack Lloyd <lloyd@acm.jhu.edu>
List: netbsd-help
Date: 03/21/2002 16:11:11
Hello,

This is my first experience using NetBSD, so forgive me if this is
something that pops up every 2 weeks.

Basically, I have a SPARCstation2 (named botan) that I am trying to boot
diskless. The server is my x86 box running Linux (named aisha). I'm using
NetBSD 1.5.2.

I've followed the NetBSD Diskless HOWTO fairly closely, setting up a root
for the machine in /usr/local/netbsd. I did differ from the HOWTO in that I
put botan's root in /usr/local/netbsd/root/botan, because I have 3 other
SS2s that I would like to set up in a similiar configuration and that way
they can share their /usr.

Anyway, I set up RARP, TFTP, NFS, etc, etc, on aisha, and when I start up
botan it works great to start out. It finds it's IP, gets the initial boot
image via TFTP, then mounts the root partition with NFS and boots the main
kernel. Bootparams seems to be working fine. At this point (right after it
scans and resets the SCSI bus) I run into trouble. It start's doing
something like (I'm not near the machine right now so this may not be
exactly right):

mount_root: timeout
mount_root: timeout
mount_root: timeout
could not mount root
And then something about a bootparam error

[Oh, I'm seeing the console over the serial line, BTW]

Anyway, using ethereal, what it looks like is that botan starts asking for
bootparam again, which my portmap then ignores. That seems like a bad
thing. But I don't get why it's asking for bootparam a second time, when it
will just be the same thing, and why it thinks the NFS mount has failed,
when it's already mounted that partition, in order to get the kernel!

Another odd thing about bootparamd. The first time it's called, it responds
with
   Returning botan   (none)    10.0.0.1
whereas the Diskless HOWTO says it should be
   Returning client   (none)    127.0.0.1

(10.0.0.1 is aisha's IP). Since I don't know squat about bootparamd, I
don't know if this is relevant or not.

So, before I really start rambling about this, does this seem more like a
problem with my portmap/bootparamd/nfs server, or something else? aisha is
a pretty stock RedHat 7.2 box, running kernel 2.4.16 and the bootparamd and
portmap that came with the distribution (there haven't been any updated
versions released). The Diskles HOWTO mentions that some Linux bootparamd's
have problems, but it says 0.17 (which I'm using) is known to work. Should
I try compiling bootparamd from source (0.17 is the latest out), or try the
portmap 5 beta? Or possibly some sort of voodoo ritual? :)

Wow, this email is way too long. Thanks for reading all the way down to
here, and for any help you might be able to give on this,

Jack